/// <summary>
/// Returns two-points defining the linear regression line.
/// </summary>
/// <param name="col">chosen column</param>
/// <param name="height">the height of the canvas</param>
/// <param name="width">the width of the canvas</param>
/// <returns></returns>
public PointCollection GetLineRegPoints(string col, double height, double width)
{
Console.WriteLine(col);
Console.WriteLine(model.getCorrelatedFeatureByFeature(col));
// calculate min & max values of correlated features
double minXVal = MinMaxVals[col][0];
double maxXVal = MinMaxVals[col][1];
double absMaxXVal = Max(Abs(minXVal), Abs(maxXVal));
double minYVal = MinMaxVals[model.getCorrelatedFeatureByFeature(col)][0];
double maxYVal = MinMaxVals[model.getCorrelatedFeatureByFeature(col)][1];
double absMaxYVal = Max(Abs(maxYVal), Abs(minYVal));
List <double> l = model.GetLineByFeature(col);
PointCollection points = new PointCollection();
// calculate the desirable ratios of x & y axes
if (absMaxXVal == 0)
{
xRegRatio = 0;
}
else
{
xRegRatio = (width / 2) / absMaxXVal;
}
if (absMaxYVal == 0)
{
yRegRatio = 0;
}
else
{
yRegRatio = (height / 2) / absMaxYVal;
}
// create two points defining the linear regression line, trying to draw in the scope of the canvas
if (minXVal > minYVal)
{
System.Windows.Point p = new System.Windows.Point(minXVal * xRegRatio + (width / 2), (height / 2) - CalcY(minXVal, l) * yRegRatio);
points.Add(p);
}
else
{
System.Windows.Point p = new System.Windows.Point((width / 2) + CalcX(minYVal, l) * xRegRatio, (height / 2) - minYVal * yRegRatio);
points.Add(p);
}
if (maxXVal < maxYVal)
{
System.Windows.Point p = new System.Windows.Point(maxXVal * xRegRatio + (width / 2), (height / 2) - CalcY(maxXVal, l) * yRegRatio);
points.Add(p);
}
else
{
System.Windows.Point p = new System.Windows.Point((width / 2) + CalcX(maxYVal, l) * xRegRatio, (height / 2) - maxYVal * yRegRatio);
points.Add(p);
}
return(points);
}
